FT : SoftBank discussed AI chips tie-up with Intel to rival Nvidia

SoftBank discussed AI chips tie-up with Intel to rival Nvidia
Son hopes to put Japanese group at centre of artificial intelligence boom but talks with US chipmaker failed to progress

SoftBank held talks with Intel about producing an artificial intelligence chip to compete with Nvidia but the plan foundered after the US chipmaker struggled to meet the Japanese group’s requirements.

Negotiations to partner with Intel, which have not previously been reported, would have accelerated SoftBank’s efforts to combine the chip designs of its crown jewel Arm with the production expertise of its latest acquisition, Graphcore, to create a rival to Nvidia’s market-leading AI chips, said people familiar with the matter.

SoftBank chief executive Masayoshi Son plans to invest billions of dollars in an attempt to put the Japanese group at the centre of the AI boom. His ambitious scheme, which he has pitched to Big Tech companies, encompasses chip production and software through to providing power for the data centres that would house its processors.

The talks with Intel failed in recent months, in advance of the US chipmaker’s announcement of drastic cost-cutting plans, including thousands of lay-offs, in early August, these people said. SoftBank is now focusing on discussions with Taiwan Semiconductor Manufacturing Co, the world’s largest contract chipmaker.

Using Intel’s US foundry to manufacture AI chips could have allowed SoftBank to tap into the Biden administration’s Chips Act funding to boost domestic semiconductor production.

Intel chief executive Pat Gelsinger is attempting to put the Silicon Valley company back on the leading edge of global chipmaking. After receiving almost $20bn in funding and loans from the US government in March, Intel is investing heavily in an attempt to catch up to rivals TSMC and Samsung in chip manufacturing and land major new customers for the company’s foundry business.

SoftBank has blamed Intel for the collapse of the talks, these people said, claiming the chipmaker was incapable of meeting its demands for volume and speed. They also cautioned the talks could start again given the limited number of chip manufacturers with the capabilities needed to produce cutting-edge AI processors.

Intel declined to comment on “discussions we may or may not have with customers”. SoftBank and Arm declined to comment.

Undeterred by the uncertainties surrounding his production plans, Son has pitched some of the world’s biggest tech groups, including Google and Meta, as he tries to drum up support and financing for his latest venture. Some of the vast investment needed to build a new chip production business could be funded by advance orders from deep-pocketed Big Tech companies, said people familiar with his thinking.

Meta declined to comment. Google did not respond to a request for comment.

One element of Son’s pitch is that SoftBank could help counter the market power of Nvidia, which briefly became the world’s most valuable company earlier this year. Nvidia’s AI data centre chips are by far the most popular on the market, with its broad software platform, Cuda, underpinning its dominance.

Critics of Son’s plan have questioned whether moving Arm into chip production could damage its relationship with Nvidia, a key client, but the people familiar with the plan say SoftBank believes the risk is worth the reward.

Son, said the same people, still intends to design and produce an AI chip, with one ambitious estimate suggesting a prototype could be ready in a matter of months. His recent purchase of struggling UK AI chipmaker Graphcore was driven by its expertise in bringing a chip into production, they added.

However, chip production capacity remains a significant hurdle. The SoftBank boss has held talks with TSMC but has not reached an agreement, as the Taiwan-based chipmaker is struggling to meet demand from existing customers, including Nvidia, the people said. TSMC declined to comment.

One of the people familiar with the still-evolving plan suggested that if an agreement with TSMC can be reached, Son could need another partner to provide the expertise in chip design that Intel had offered.

The cost of Son’s latest venture could run into the tens of billions of dollars but people close to SoftBank cautioned that putting a figure on the total investment needed is not realistic at this stage. The chief executive has sounded out Saudi Arabian and United Arab Emirates investors about the plan but nothing has been agreed, they said.

Intel, which had been a cornerstone investor in Arm’s initial public offering last September, disclosed this week that it sold its entire stake in the UK chip designer during the second quarter of this year, raising about $150mn. The company recently suspended its dividend as it tries to conserve cash.

In April, Intel revealed a $7bn operating loss for its manufacturing business, causing shares to plunge. Reports of a design flaw in its PC chips followed. Then, during its latest results announcement earlier this month, it launched a plan to cut about 15 per cent of its workforce, amid declining revenue. Its shares shed a quarter of their value in a day, leaving its market capitalisation languishing below $100bn.

FT : How Roche passed on a potential $14bn-a-year weight-loss pill

How Roche passed on a potential $14bn-a-year weight-loss pill
The Swiss company had the first option on a drug that could now become a leading anti-obesity treatment

Swiss drugmaker Roche turned down a potential $14bn-a-year weight-loss pill now being developed by rival Eli Lilly that could have given it a leading position in the booming weight-loss market, according to company filings and people close to discussions.

In 2018, three years before Novo Nordisk’s’ Wegovy was approved as the first weekly anti-obesity medication, Roche opted not to take advantage of its “right of first refusal” to purchase a type-two diabetes pill from Japanese drugmaker Chugai, with which it has a two-decades long partnership.

The drug — then known as Owl-833 — was ready to enter early phase-one trials and was valued at just tens of millions of dollars, two people familiar with the matter said. It was instead bought by Eli Lilly that year for an upfront payment of $50mn and became orforglipron.

The daily weight-loss pill is still in clinical trials but could be approved by 2026 and is expected to generate $50bn in global revenues for the Indianapolis-based drugmaker in the six years after its launch, reaching $14.4bn in annual sales in 2032, according to analyst consensus estimates.

“Roche passed up on a mega-blockbuster,” said David Risinger, a biopharma analyst at Leerink Partners.

How Roche missed out on the Eli Lilly drug — orforglipron — before eventually re-entering the weight-loss drug race five years later with its acquisition of Carmot Therapeutics shows the challenges of developing anti-obesity medications and how big pharmaceutical groups are now scrambling to buy back into the field.

Pharma groups have a history of passing on early-stage drugs which go on to become very lucrative, noted Louise Chen, an analyst at Cantor Fitzgerald.

In a recent example, Pfizer partially licensed an experimental gut disease drug to biotech Roivant Sciences in late 2022 for free. In less than a year, Roivant had sold the drug to Roche for $7bn.

Eli Lilly’s decision to take a bet on orforglipron and fund its development shows how the biggest players in insulin-based diabetes treatments “had an edge in understanding metabolism and weight loss,” said a person close to the licensing discussions in 2018.

“Either you have the focus and expertise inside a company or you don’t, it’s not like drugs land on your desk and become a blockbuster.”

Roche has recently re-entered the weight-loss market by buying anti-obesity drug developer Carmot for up to $3.1bn last year.

Since then, it has released early data from both an injectable and a weight-loss pill, as it bids to challenge the leading positions of Novo Nordisk and Eli Lilly, whose weight-loss injectables — Wegovy and Zepbound — are already generating billions in sales.

Companies are now vying to develop an easily scalable weight-loss pill, which would be easier to administer, in the next step to compete in the the $130bn a year weight loss market.

Weight-loss pills activate similar gut hormone receptors, such as GLP-1, to the existing injectable medications, helping to suppress appetite. Analysts believe Eli Lilly is ahead in this race, thanks in large part to orforglipron.

The pill is expected to launch in 2026 and will be the only small molecule GLP-1 drug — a type of medicine based on a simple chemical formula easily reproducible as a pill — available for at least two years, giving Eli Lilly the chance to “blanket the world” with the medicine, added Risinger.

Lilly is already investing in expanding manufacturing capacity to cope with the potential huge demand for its weight-loss pills. In a mid-stage trial, orforglipron cut participants’ body weight by up to 14.6 per cent.

Rival drugmaker Novo Nordisk’s weight loss pill, oral semaglutide, is more complex to manufacture, likely hampering its rollout.

Early results from a Roche small molecule weight-loss pill last month gave users a 6.1 per cent reduction in weight compared to a placebo, leading to excitement among investors.

Roche’s weight-loss drugs, meanwhile, are not expected to launch before 2028. Roche is set to generate just $4bn in sales from its weight-loss injectable and pill by 2032, according to analyst projections.

However, the drugs’ patents may infringe on rival medicines, according to a regulatory filing submitted by Carmot before it was acquired.

Roche’s business development team assessed the early version of orforglipron in 2018, but concluded that “it wasn’t a priority”, according to a person close to discussions.

Roche had sold an earlier obesity drug — Orlistat — in the 2000s, as it caused gastrointestinal side effects.

After a GLP-1 drug known as taspoglutide flopped in late-stage trials in 2010, Roche reduced its investment in diabetes and obesity drugs, focusing instead on its cancer pipeline.

Gareth Powell, head of healthcare at UK investor Polar Capital and a backer of weight-loss company, Zealand Pharma, said: “There wasn’t the same excitement about weight-loss then that there is now. You have to respect [Roche] for seemingly admitting they were wrong and going after it again.”

Roche said: “Healthcare R&D is always associated with risks, prioritisation and robust decision making on where to focus investments so that they can have the biggest potential impact on patients.

“Portfolio decisions follow these principles. In 2018, neither the broader potential application of [GLP-1s] was clear yet nor was the scientific knowledge advanced as it is today.”

Eli Lilly and Chugai declined to comment.

Roche has also had big wins from its partnership with Chugai: a $4.5bn-a- year haemophilia treatment Hemlibra, one of Roche’s biggest drugs, was licensed from the Japanese drugmaker.

Chugai, in which Roche holds a nearly 60 per cent stake, is still set to benefit from orforglipron’s success, as it will make up to $390mn based on achieving regulatory and sales milestones.

FT : Germany’s steelmakers are caught in a tightening trap

Germany’s steelmakers are caught in a tightening trap
The sector’s green transformation is a threat to competitiveness

German steelmakers are beset with problems. Sluggish demand, rising Chinese imports and falling prices are depressing their current performance. The sector’s green transformation, instead of offering a brighter future, poses another threat to their long-term competitive position. 

For evidence of this tightening steel trap, look no further than Thyssenkrupp. It is embroiled in the complex carve-out of its steel unit, in which Czech entrepreneur Daniel Křetínský has bought a 20 per cent stake and is in talks over a further 30 per cent. The unit is performing poorly: operating profits nearly halved in the third quarter. It faces overcapacity as its core German automaker client base loses market share. Thyssenkrupp is involved in a tussle with the steel unit’s management over how much of a dowry to provide it. 

Meanwhile, Thyssenkrupp Steel has undertaken to switch from coal to renewable hydrogen. This is a costly proposition.

The capex involved in converting plants is manageable, perhaps $800mn for a 2mn tonne direct reduction (DRI) plant according to analysis by Rafal Malinowski of the Energy Transitions Commission, which works out at $43 per tonne of steel produced. The real problem is the cost of green hydrogen needed to feed the DRI plant. Price discovery for Thyssenkrupp is ongoing: it has put out a tender for green hydrogen supply. But a ballpark figure for the energy required to make one tonne of steel might be $450. Plug in the cost of ore, labour and everything else the plant needs — at some $380 per tonne — plus capital amortisation, and you get a cost of steel of $870/tonne.


Thyssenkrupp has been awarded €2bn in capex and opex support to plug the gap with the cost of making dirty steel, which is today $580 per tonne according to Tom Zhang at Barclays, rising to perhaps $730 per tonne by 2035. Over time, too, the “green premium” should shrink as carbon prices continue to rise and hydrogen production technologies scale up.

But the more pertinent comparison should not be with Europe’s dirty steel, with which the continent’s green version will eventually become competitive. It should be with green steel produced in areas of the world that benefit from cheaper renewables, such as the Middle East.

The move from using near-field coal to faraway renewables puts German steelmakers at a long-term, structural disadvantage. After all, moving steel across the water will always be cheaper than shipping hydrogen. Such a gap cannot easily be plugged. While decarbonisation, employment and strategic independence may be desirable political objectives, “forever” subsidies will prove a high cost to pay.

FT : Kamala Harris should not cave to overtures from the crypto crowd

Kamala Harris should not cave to overtures from the crypto crowd
Industry lobby groups are seeking the lighter touch of the CFTC regulation rather than SEC oversight

After years of crypto kingpins being handcuffed and sent to prison, numerous spectacular bankruptcies, rampant fraud and manipulation, breathtaking volatility and a long list of lost court cases, the crypto industry is nonetheless riding high in the US.

That’s partly because it has a huge cash pile that it is willing to spend on campaigns to buy the support of politicians who will back its special interest agenda. The crypto industry’s big goal is to pick its own regulator and get a veneer of legitimacy, but not be regulated much at all.

As the Securities and Exchange Commission is a very powerful and effective cop on the crypto beat, the industry sees this regulator as its “mortal enemy”. The crypto crowd wants its political allies to put the smallest, least funded, least capable, and most easily capturable financial regulatory agency in charge of crypto — the Commodity Futures Trading Commission.

With crypto, it is clear from many cases that almost all of the tokens traded comfortably fall within the standard definition of securities and should be regulated by the SEC as such. Those that aren’t securities comfortably fall within the standard definition of commodities and should be regulated by the CFTC as such.

There is really very little dispute about this among people who are not on the payroll of the crypto industry. And that’s also why the SEC is winning almost all the legal cases it is bringing against crypto companies, which argue that most if not all the securities, commodities and banking laws that apply to every other financial firm in America don’t apply to them.

Less than two years after numerous politicians were scrambling to return industry campaign contributions from the fraud-filled FTX, crypto is emboldened to the point it is setting its sights on influencing the Kamala Harris campaign for president. One reported argument is the supposed need to counter Donald Trump’s embrace of crypto.

The crypto industry appears to be making some headway. Officials from the Biden administration and the Harris campaign recently held a conference call with industry figures. Harris should reject the overtures. Here’s why:

First, after years of effort and claims that cryptocurrencies have a real value, there is still no real case to use them for legitimate purposes over existing currencies. They remain the financial product of choice among financial predators, lawbreakers and criminals worldwide. The least harmful use is wild speculation and gambling (as opposed to its other uses for tax evasion, fraud, ransomware, sanctions evasion, terrorist funding, narcotics trafficking, money laundering, etc.).

Second, easing crypto regulation is not among the top concerns of the American people. Contrary to industry propaganda, only about 18mn adult Americans even use or own crypto and that number is declining, according to Federal Reserve survey data.

It really is a very niche issue. Of the 88 per cent of Americans who have heard about crypto, a Pew Research survey last year found a supermajority of 75 per cent are not confident or not very confident about the reliability and safety of cryptocurrencies. Importantly, between 61 and 77 per cent of voters in six key swing states have a negative view of crypto, according to the venture capital firm Digital Currency Group and the polling firm Harris Group (unrelated to the vice-president).

Third, the crypto industry’s extensive lawbreaking rap sheet is at odds with Harris’s long and strong record as a prosecutor who fights for consumer and investor protections and against financial industry lawbreaking. Remember, when she was California’s attorney-general, she was under enormous pressure to accept a global subprime mortgage settlement with Wall Street’s biggest, most powerful banks. Harris was tough, reportedly even saying no to JPMorgan’s chief executive Jamie Dimon over a settlement. That’s not easy. But she held firm and cut a much better deal for California.

Finally, communities of colour are disproportionately victims of crypto rip-offs. Yes, these communities are rightly sceptical of the traditional financial system that has excluded, discriminated and exploited them for so long. Unfortunately, that makes them a target for the crypto industry, which pitches bogus wealth-building opportunities. A 2021 survey by the social science research institute NORC at the University of Chicago estimated 44 per cent of crypto traders were not white.

Harris has a lot to do in the lead-up to the US elections. Caving to threats from the crypto industry should not be one of them.

>>> Paulson & Co (John Paulson) discloses updated portfolio positions in 13F fi

Paulson & Co (John Paulson) discloses updated portfolio positions in 13F filing: Added to MDGL holding, Exited NMRK TAST BABA
Highlights from Q2 2024 filing as compared to Q1 2024 (all amounts are approximate):
  • Increased positions in: MDGL (to 1.85 mln shares from 1.78 mln shares)
  • Maintained positions in: THM (64.2 mln shares), NG (27.24 mln shares), BHC (26.44 mln shares), PPTA (24.77 mln shares), BSIG (8.95 mln shares), SA (2.07 mln shares), THRY (2 mln shares), AEM (0.78 mln shares)
  • Closed positions in: NMRK (from 3 mln shares), TAST (from 0.06 mln), BABA (from 0.05 mln)
  • Decreased positions in: AU (to 1.93 mln shares from 2.93 mln shares)

>>> US After Hours Summary: CSCO +5.6% higher on earnings and restructuring; LIT

After Hours Summary: CSCO +5.6% higher on earnings and restructuring; LITE +18.2% higher on earnings; ULTA +14.5% surging as Warren Buffett discloses new position; NKE +4.2% as Bill Ackman discloses new position

After Hours Gainers:

Companies trading higher in after hours in reaction to earnings/guidance: LITE +18.2%, ARDT +9.1%, PYCR +6.5%, NVGS +6.4%, CSCO +5.6% (also announces restructuring including 7% workforce reduction), STNE +0.4%

Companies trading higher in after hours in reaction to news: ULTA +14.5% (Berkshire Hathaway (Warren Buffett) discloses new position), SIRI +5.8% (Berkshire Hathaway (Warren Buffett) discloses increased position), NKE +4.2% (New position from Pershing Square (Bill Ackman)), RAMP +3.6% (increases share repurchase authorization by $200 mln to $1.3 bln), HEI.A +3.3% (Berkshire Hathaway (Warren Buffett) discloses new position), WRBY +2.3% (Director bought 75,000 shares), CIEN +2% (in sympathy with CSCO earnings), HPE +1.8% (in sympathy with CSCO earnings), IMMR +1.7% (to delay 10-Q filing), RDW +1.5% (to acquire Hera Systems; raises FY24 rev guidance), PEB +1.2% (Chairman/CEO bought 30,000 shares), PSN +0.7% (unveils updated iNET), HOOD +0.6% (July operating data), NVST +0.6% (CEO bought 30000 shares), EXTR +0.4% (in sympathy with CSCO earnings), ICLR +0.4% (names new CFO), VREX +0.3% (CFO bought 10000 shares)

After Hours Losers:

Companies trading lower in after hours in reaction to earnings/guidance: TITN -21.9%, CWCO -5.9%, ARQT -4.5%, PGEN -3.6%, CELC -2% (also now expects to reach the enrollment target later than expected), ASTS -2%, DLO -0.6%

Companies trading lower in after hours in reaction to news: SPIR -18.3% (to delay 10-Q filing), NHI -4% (files supplement for 2 mln share offering), FIHL -2.9% (authorizes new $200 mln share repurchase program, also reports earnings), NEXT -1% (provides Q2 business update), DOV -0.5% (launches its new EvoDose NXT Chemical Dispenser), AVGO -0.1% (in sympathy with CSCO earnings), MARA -0.1% (completes $300 mln convertible notes offering)

>>> SRS Investment discloses updated portfolio positions in 13F filing: New HPE

SRS Investment discloses updated portfolio positions in 13F filing: New HPE TXG UAL UNH RBLX positions, Added to SNAP, Exited AMD DDOG MGM, Lowered AAL
Highlights from Q2 2024 filing as compared to Q1 2024 (all amounts are approximate):
  • New positions in: HPE (3 mln shares), TXG (0.5 mln), UAL (0.49 mln), UNH (0.3 mln), RBLX (0.07 mln)
  • Increased positions in: SNAP (to 72.72 mln shares from 65.11 mln shares), BURL (to 1.01 mln from 0.21 mln), TEAM (to 1.35 mln from 1.08 mln)
  • Maintained positions in: CAR (17.13 mln shares), PLNT (6.11 mln shares), PDD (3.65 mln shares), NFLX (2.14 mln shares), NVDA (1.7 mln shares), META (1.06 mln shares), TSM (0.77 mln shares), SPOT (0.4 mln shares)
  • Closed positions in: AMD (from 2.16 mln shares), DDOG (from 1.48 mln), MGM (from 0.28 mln), LAD (from 0.14 mln), PYPL (from 0.14 mln)
  • Decreased positions in: AAL (to 13.06 mln shares from 21.64 mln shares), GLBE (to 0.16 mln from 0.63 mln)

>>> Glenview Capital (Larry Robbins and Mark Horowitz) discloses updated portfol

Glenview Capital (Larry Robbins and Mark Horowitz) discloses updated portfolio positions in 13F filing: New ZI CNC BHC LYFT WCC MTCH LYV positions
Highlights from Q2 2024 filing as compared to Q1 2024 (all amounts are approximate):
  • New positions in: ZI (1.57 mln shares), CNC (1.43 mln), BHC (1.31 mln), LYFT (0.82 mln), WCC (0.38 mln), MTCH (0.35 mln), LYV (0.23 mln), CRS (130K), TFX (122K)
  • Increased positions in: CVS (to 9.16 mln shares from 1.7 mln shares), ALIT (to 30.86 mln from 28.32 mln), DNA (to 26.5 mln from 24 mln), TEVA (to 9.16 mln from 7.5 mln), VTRS (to 6.49 mln from 5.14 mln), CLVT (to 25.65 mln from 24.65 mln), GPN (to 3.12 mln from 2.39 mln), EVLV (to 6.55 mln from 5.95 mln), EXAS (to 0.63 mln from 0.23 mln), CAR (to 0.53 mln from 0.18 mln), OFIX (to 1.11 mln from 0.88 mln), KNX (to 1.45 mln from 1.33 mln), UBER (to 1.2 mln from 1.13 mln)
  • Maintained positions in: HCA (0.2 mln shares)
  • Closed positions in: ARRY (from 2.2 mln shares), W (from 220K), LBTYA (from 176K), CBUS (from 117K), META (from 110K), AVGO (from 20K)
  • Decreased positions in: BKD (to 7.01 mln shares from 14.44 mln shares), DXC (to 8.71 mln from 12.78 mln), ESI (to 6.33 mln from 8.36 mln), CTVA (to 5.07 mln from 6.79 mln), THC (to 4.67 mln from 6.33 mln) EXPE (to 600K from 614K), USFD (to 2.49 mln from 2.52 mln), MYGN (to 3.75 mln from 4.94 mln), VVV (to 0.48 mln from 1.63 mln), UHS (to 1.18 mln from 2.05 mln), FI (to 0.05 mln from 0.8 mln), MRVL (to 0.86 mln from 1.37 mln), CI (to 1.29 mln from 1.47 mln), RPID (to 1.2 mln from 1.32 mln), ELV (to 0.1 mln from 0.17 mln), BC (to 426K from 481K), AMZN (to 350K from 403K), MCK (to 266K from 288K)

>>> Viking Global (Andreas Halvorsen) discloses updated portfolio positions in 1

Viking Global (Andreas Halvorsen) discloses updated portfolio positions in 13F filing: New F DASH UBS MCD INSM MMM NVDA DDOG NFLX positions, Added to AMZN DLTR EQH LVS USB, Exited CNH V KDP AMD
Highlights from Q2 2024 filing as compared to Q1 2024 (all amounts are approximate):
  • New positions in: F (18.79 mln shares), DASH (5.54 mln), UBS (5.19 mln), JPM (3.01 mln), APO (0.97 mln), MCD (1.88 mln), INSM (1.04 mln), MMM (1.64 mln), NVDA (1.38 mln), DDOG (1.38 mln), NRG (0.94 mln), NFLX (0.93 mln), RBRK (0.9 mln), CVNA (0.55 mln), ALL (0.55 mln), SPOT (0.47 mln), CHTR (0.44 mln)
  • Increased positions in: DLTR (to 6.81 mln shares from 2.57 mln shares), EQH (to 5.86 mln from 2.06 mln), LVS (to 19.79 mln from 16.92 mln), HDB (to 8.76 mln from 5.93 mln), USB (to 25.4 mln from 22.63 mln), AMZN (to 9.29 mln from 6.9 mln), CPT (to 3.26 mln from 1.03 mln), ADBE (to 2.28 mln from 0.56 mln), OSCR (to 5.93 mln from 4.73 mln)
  • Maintained positions in: BBIO (25.12 mln shares)
  • Closed positions in: CNH (from 11.96 mln shares), VSTM (from 2.46 mln), V (from 6.33 mln), BEKE (from 5.25 mln), KDP (from 4.46 mln), AMD (from 3.95 mln), DHR (from 2.84 mln), VVV (from 2.38 mln), KNTE (from 1.82 mln), TTWO (from 2.01 mln), BSX (from 1.72 mln), KNX (from 1.72 mln), CBOE (from 1.66 mln), LW (from 1.21 mln), SKX (from 1.09 mln), MA (from 0.79 mln), PPG (from 0.62 mln), HUM (from 0.46 mln)
  • Decreased positions in: FCX (to 5.23 mln shares from 9.52 mln shares), PM (to 10.91 mln from 13.02 mln), APG (to 27.03 mln from 28.98 mln), TMUS (to 2.49 mln from 4.38 mln), EWTX (to 1.48 mln from 3.34 mln), FIVN (to 1.73 mln from 3.45 mln), WDAY (to 3.96 mln from 5.29 mln), CBRE (to 3.62 mln from 4.57 mln), UNH (to 0.44 mln from 1.36 mln), AMP (to 0.91 mln from 1.63 mln), KKR (to 0.55 mln from 1.07 mln), META (to 0.27 mln from 0.73 mln)